Martin P.M. Handbook of Deposition Technologies for Films and Coatings, Third Edition: Science, Applications and Technology

Подождите немного. Документ загружается.

Characterization of Thin Films and Coatings 771

Exhaustive and quantitative description of the physics behind the Raman effect and the

analysis techniques is outside the scope of this handbook, hence the interested reader is

advised to follow the references and various information sites that result from a web search.

Applications

RS is primarily a non-destructive, structural characterization tool. The vibrational frequencies

are sensitive to bond characteristics such as length, strength, and their arrangement. Mass,

equilibrium spatial arrangement, relative displacement during vibration, and bond force

constants of atoms in a molecule or crystal can inﬂuence the frequency of the Raman peak

[37]. Therefore, RS can be wisely used for understanding defect structure and disorder in

crystals coupled with the chemical information obtained from other techniques such as FTIR.

Although the technique was initially developed for understanding the scattering of light in

various liquids, currently minerals, powders, thin ﬁlms, organic/inorganic coatings, and single

and polycrystalline materials are all subjects of investigation. RS is used for ﬁngerprinting,

understanding vibrational motions of crystals with the help of selection rules, phase

transformations, defect structure and disorder analysis, structure of amorphous gels and

glasses. Raman microscopy is effectively used for studying the inclusions in grain boundaries,

fracture analysis, precipitate and interfacial analysis. In the backscattering geometry and

cross-sectional sample-view, coatings and multilayers can be analyzed by avoiding any

scattering from the substrate [6]

Phase transformations in crystalline, inorganic thin ﬁlms and coatings

Vacuum and ambient annealing treatments are carried out to understand the thermal stability

of coating materials such as TiN, CrN, and TiAlN, and their multilayers [38, 39]. The

variations in the spectral positions and the intensity of the optical phonon mode have been

used to understand the phase transformations and the defect structure in the coatings. Barshilia

and Rajam [38] used Raman analysis to precisely determine the oxidation onset in various

materials and reported that the TiAlN/CrN multilayered coatings are stable until 900

◦

C. It was

reported that the micro-Raman analysis could be used to detect the presence of minor phases

such as anatase-TiO

which are difﬁcult to trace by XRD. Micoprobe Raman analysis has been

used to detect the presence and spatial distribution of the anatase and rutile TiO

phases and

the stoichiometric changes in the TiO

was determined by studying the Raman active peak

shifts due to oxygen deﬁciencies. Hong et al. [40] used RS to study the effect of annealing

temperature on TiO

phase evolution. As shown in Figure 16.13(b), they observed a gradual

transformation of as-deposited amorphous ﬁlm to crystalline TiO

ﬁlm with increasing anatase

to rutile ratio with increasing temperature. Inorganic coatings (hydroxyapatite (HA),

tricalcium phosphate (TCP)) on biomedical implants (Ti-based) have also been analyzed with

the help of micro- and macro-Raman analyses. The key advantage of Raman analysis is its

ability to distinguish HA and TCP more easily than XRD and FTIR [41].

772 Chapter 16

Figure 16.14: Depiction of the sample orientations in plane (left) and cross-section (right)

geometries indicating the advantages of cross-section geometry in avoiding substrate signal. In

the ﬁgure, while the region indicated by black arcs shows the interaction area different arrows

indicate the depth from which the signal is obtained.

Depth proﬁle and cross-sectional analysis using μ-Raman spectroscopy

Raman microscopy or μ-RS is not just conﬁned to the surface but can be effectively used for

studying the depth distribution. By changing the energy of the incident light the penetration

depth can be varied and the desired depth resolution can be obtained. However, this technique

suffers from an inherent difﬁculty if the penetration depth of the material (coating or ﬁlm) is

much higher than the thickness because the substrate peaks also contribute to the spectra,

thereby spoiling the resolution. This problem becomes more severe when the analysis is

desired on multilayered and transparent ﬁlms. However, if the desired layers are the size

of the optical beam or larger, samples can be analyzed in the cross-sectional geometry

(Figure 16.14). As opposed to the plan-view geometry where all the participating layers

contribute to the phonon characteristics, the cross-sectional geometry with linear scans at

multiple positions can provide a detailed picture of the multilayered heterostructures. BN, SiC,

carbon, diamond and diamond-like carbon (DLC) ﬁlms on Si substrates were analyzed by

Werninghaus with cross-sectional geometry [42].

Organic/polymer coatings and corrosion science

One of the most important applications of RS in coatings technology is the compositional

mapping of coatings. Sequential acquisition of Raman spectra by incrementally rastering the

laser beam across the sample can yield a compositional map. Based on the number of Raman

bands to be tracked, the complexity and acquisition times will increase. Line scans are also

possible with the help of a charge-coupled device (CCD) camera and a spectrograph. Use of a

water-immersion objective for studying ligand distribution on chromatographic adsorbent

particles was reported by Larsson et al. [43]. The ability of this technique to distinguish the

core and coatings was used to identify the uniform distribution of allyl groups and surface

layers of sulfopropyl groups. The surface morphology of rubber domains within the coated

Characterization of Thin Films and Coatings 773

substrate contributes strongly to the material characteristics such as adhesion and impact

resistance. While conventional imaging analyses were unsuccessful in studying the interfacial

characteristics, the use of Raman imaging coupled with contrast enhancement functions

resulted in a better understanding of near-surface composition of thermoplastic oleﬁn (TPO)

coated with chlorinated polyoleﬁn [44]. Owing to the excellent spatial resolution available

(∼1 ␮m) from Raman imaging, it can be used for studying the crystallinity gradients of the

polymer coatings on metal substrates and to evaluate the mechanical performance. The

use of polarized confocal Raman microscopy to track the polymer molecular orientation

has also been reported. The combination of confocal laser scanning microscopy and

ATR-FTIR with confocal Raman microscopy was used to track the UV curing of polymer

coatings [41].

Corrosion is a widely studied problem and Raman microscopy can uniquely contribute to this

ﬁeld by in situ monitoring of the degradation. The laser can be focused onto the metal surface

through the surface coating and the spectra can be collected as a function of time. Multiple

forms of iron oxides and hydroxides have been identiﬁed with the help of Raman analysis of

coated steel coupons immersed in a corrosive environment. Redox states of the protective

coating and the solution/coating interfacial product analysis are key advantages of RS in the

ﬁeld of corrosion science. Surface-enhanced Raman scattering (SERS) has been found to be

very useful in analyzing the anticorrosion coatings on Cu surface [41]. In principle, the SERS

technique uses a large increase in the electric ﬁeld strength near a rough (noble) metal surface

coupled with a chemical enhancement due to a charge-transfer between the coating (adsorbate)

and the metal. This method requires careful attention to a number of parameters such as

surface roughness, reactivity of metal and adsorbate under the laser beam, and further

complications are also possible owing to complex spectra with numerous other features due to

selective band enhancement and molecular orientation effects on the surface.

Recently, scientists from NIST [45] have reported the use of coherent anti-Stokes Raman

scattering (CARS) as a possible alternative for SERS. The coherent nature of CARS helps

improve the collection efﬁciency and sensitivity. This technique also eliminates the need for

coinage metals (Ag, Au, and Cu), improves the signal-to-noise ratio, and provides the ability

analyze the inﬂuence of interaction into the bulk of the substrate as well. This technique, when

completely developed, is expected to be used for the characterization of advanced electronic

applications such as organic light-emitting diodes and thin ﬁlm transistors.

As evident from the various examples above, Raman microscopy is a very important tool for

the characterization and property evaluation of thin ﬁlms and coatings in a wide range of

applications.

Strengths:



High sensitivity to the chemical bonds and their variations.

774 Chapter 16



Offers suitable spatial resolution for compositional mapping of coatings and

multilayer heterostructures (regions as small as ∼1 ␮m in diameter).



Provides the ability to study phase (material) transformations in situ for liquid, solid

(organic and inorganic), thin ﬁlms and coatings.



Flexible sample handling.



Signals can be enhanced using various methods, e.g. resonance Raman enhancement

or SERS.

Limitations:



Very low-intensity signal – partly solved by using high-intensity lasers and specialized

equipment – however, some instruments and techniques impose various other

restrictions.



A common source of interference is ﬂuorescence or other photoluminescence from the

sample.



Raman intensities and bandwidths are also dependent on crystallinity. Features from

amorphous materials can be weak and broad.



Absolute quantiﬁcation may not be possible.



Difﬁcult to understand and track trace chemical impurities.

16.3.3.2 Infrared Spectroscopy

S. V.N.T. Kuchibhatla

IR spectroscopy is the subset of optical spectroscopy that uses the IR portion of the spectrum,

the central part of which ranges from less than 2 ␮m to more than 30 ␮m in wavelength. For

most of this frequency range the photon energies (0.6–0.04 eV) are not usually large enough to

excite electronic levels, but may excite vibrational or rotational states of covalently bonded

atoms or groups. Because molecules rotate or vibrate at speciﬁc frequencies it is possible to

identify molecular groups and many of the types of molecular bonds present, provided they

satisfy the dipole selection rule.

The Fourier transform (FT) of IR spectroscopy (making it FT-IR or FTIR) involves the use of a

beam splitter, movable mirrors, a multiwavelength IR source and a Michelson interferometer,

as shown in Figure 16.15. Because IR light is usually of low intensity, the use of FT

spectroscopy [46] allows the collection of better data in a shorter time. The data collected are

actually an interferogram that is the FT of the absorption (or reﬂection spectra).

Characterization of Thin Films and Coatings 775

Figure 16.15: Schematic drawing of FTIR spectrometer set up for analysis of signal reﬂected from

a sample.

FTIR can be used to identify the molecular structure of organic contaminants, to identify the

presence of organic layers or particles, and with careful use the thickness can be determined

[47]. Because FTIR can be done in air or other environments, it is possible to monitor sorption

or decomposition processes as they occur. Spectra can be obtained in either transmission or

reﬂectance modes. Signals from thin ﬁlms can be compared and differences quantiﬁed. FTIR

can be used to identify or verify ﬁlm thickness and composition; it is also useful for examining

surface residues and identifying contaminants.

Applications

FTIR has been useful for determining the thickness of epitaxially grown ﬁlms of Si [48, 49]

and other advanced thin ﬁlm materials for which chemical bonding information, ﬁlm thickness

(including ultrathin oxides and photoresists), and dielectric function are important [50]. Also,

FTIR is used to characterize polymer coating materials, for the identiﬁcation of microscopic

defects during weathering and curing processes. As mentioned before, FTIR is often used as a

complementary technique along with RS. While Raman spectra are used to track C

consumption and O

O production, FTIR is used to detect O O H production during the

thermal/UV cure of polymer coatings. Agbenyega et al. [51] have listed the comparative

strengths and weakness of signals from various bonds from Raman and FTIR spectra. FTIR in

attenuated total reﬂection (ATR) geometry, known as ATR-FTIR, can offer a depth resolution

up to 2 ␮m, making it a more surface-sensitive technique. Hence, ATR-FTIR is widely used to

study surface degradation during weathering of polymer coatings. Ellis et al. [52] reported the

use of ATR-FTIR and FT-RS for studying changes in polyurethane-acrylic paints (containing

pigments with titanium and iron oxides) on phosphated-steel panels. It was shown that the

776 Chapter 16

Figure 16.16: Red pigmented panels, fresh and weathered samples. (a) FT-Raman and (b)

ATR-FTIR spectra [52].

ATR-FTIR is better suited for understanding transitions in the urethane bands. Also, in the case

of panels with very high absorption such as red colored panels with pigments containing iron

oxide, the FT-Raman signal is very weak. However, major changes have been observed in the

urethane band (1525 nm

−1

) region of ATR-FTIR spectra. FT-Raman and ATR-FTIR spectra

corresponding to weathered and unweathered conditions are shown in Figure 16.16(a, b) [52].

Strengths:



Identiﬁcation of organic groups and compounds.



Can operate in ambient or other environmental conditions.



Can identify some inorganic compounds.



Usually non-destructive.

Characterization of Thin Films and Coatings 777

Limitations:



Limited sensitivity.



Limited inorganic information.



Quantitative with standards or in special cases.



Minimum analysis area ∼10 ␮m.

16.3.4 Ellipsometry

M.H. Engelhard

Ellipsometry is a century-old optical method which, with modern approaches of computer

control and data processing, is becoming increasingly important for the analysis of many types

of thin ﬁlms [53, 54]. The technique is based on the analysis of polarized light reﬂected from

the sample which probes the complex refractive index (or dielectric function tensor) which can

be used to extract information about ﬁlm thickness, roughness, crystal quality, chemical

composition, or electrical conductivity. The material is probed to the optical depth of the

sample. It is sensitive enough to measure properties of ﬁlms as thin as a few tenths of

nanometers, but, if the optical properties allow, it can be used to study layers as thick as several

micrometers.

When light is incident on the surface, some of the light is reﬂected from the surface and some

light enters the material, as shown in Figure 16.17. The portion of light entering the material

does not continue at the sample angle, but is refracted according to Snell’s law. As shown in

Figure 16.17, the electromagnetic ﬁeld of the incident light can be broken into two

components, one in the plane of incidence to the sample (p) and the other perpendicular (s).

Based on the properties of the material, the attenuation and phase shift of these

electromagnetic components differ. Ellipsometry typically reports the differences in

reﬂectivity using two parameters to record the differences in the amplitude and phase of the

reﬂected waves using the equation:

ρ = R

= tan ψ exp (i) (16.3)

In this equation delta () indicates the difference between the phase angle of the parallel and

perpendicular components of the incoming wave and outgoing wave, and psi (ψ) is the angle

778 Chapter 16

Figure 16.17: Schematic drawing of incident and reﬂected components of a monochromatic beam

of light interacting with a surface. The electric vectors of the plane polarized light are labeled as p

waves (in the place of incidence) and s waves (perpendicular to the plane of incidence). Note that

a portion of the light enters the material.

whose tangent is the ratio of the magnitudes of the total reﬂection coefﬁcients. These values

are related to the ratio of the Fresnel amplitude reﬂection coefﬁcients for the p- and s-polarized

light.

Single wavelength measurements at one angle can be used to determine ﬁlm and sample

properties. However, additional information can be available if measurements are taken as a

function of wavelength (or angle) which allows the complex refractive index of the

sample to be obtained from which a number of fundamental physical properties can be

extracted [55].

Information about the sample requires the use of models to ﬁt the ellipsometric data. Typically

this involves: (1) creating a model that represents the sample, including the number of layers

present; (2) associating a dispersion relation of the optical properties of each layer; and (3)

ﬁtting the experimental data to the model. Many of the recent advances in ellipsometry involve

the use of computers and relate experimental data to appropriate models [56].

Ellipsometry measurements are widely used to determine the thickness of surface coatings and

thin ﬁlms. As already noted, the absolute thickness depends on the model and parameters used

in the model. As an example, the thickness of two NIST standard thin SiO

ﬁlms on Si (#2531)

was measured by ellipsometry and XRR and the results compared. The NIST reference values

for  and ψ are 90.03 ±0.04

◦

and 22.18 ±0.04

◦

for one sample (sample L) and 89.79 ±0.04

◦

and 22.32 ±0.04

◦

for the other (sample R). These were obtained at a wavelength of 632.8 nm

and at an angle of incidence φ of 71.330 ±0.004

◦

. Using a single-layer model this translates

Characterization of Thin Films and Coatings 779

into an oxide layer thickness of 47.7 ±0.5 nm for sample L and 48.0 ±0.5 nm for sample R.

We used a different type of ellipsometer that collected a range of wavelengths and obtained the

thickness as 47.89 ±0.5 nm for sample L and 48.15 ±0.77 nm for sample R. These results are

well within the stated uncertainty for these measurements. For comparison we also conducted

XRR on these samples and obtained thickness values of 47.6 nm for sample L and 48.0 nm for

sample R. These measurements clearly demonstrate that the thickness values measured by

ellipsometry and XRR are within the expected uncertainty of the measurements.

In another example, ellipsometry was used to measure the variation of oxide thickness over the

surface of an oxidized Si wafer. This wafer was planned to be used as a reference for

determining ion sputter rates and ion sputter rate consistency. Although many wafers show

excellent uniformity, the example shown in Figure 16.18 shows signiﬁcant variation in oxide

thickness with the difference between the minimum and maximum values of approximately

20 nm for an oxide of nominal thickness of 40 nm.

Strengths:



Highly sensitive, rapid and reproducible, comparison measurement.

Figure 16.18: Drawing of oxidized Si wafer divided into 49 coupons to use as sputter proﬁle

reference materials. The thickness (in nm) of each coupon was measured by ellipsometry using

the one-layer model, as marked on the drawing. For this particular wafer both ellipsometry and

XRR indicate that there are signiﬁcant variations in the oxide thickness depending on the location

on the wafer.

780 Chapter 16



No reference measurement needed (but knowing some parameters helpful or

necessary).



Many important properties of materials can be inferred, including both real and

imaginary parts of the dielectric function.



Can be conducted during growth and processing if optical path available (no vacuum

requirement).

Limitations:



Model dependent for physical properties.



Special approaches are needed for thick ﬁlms and thick multilayers.

16.3.5 Photoluminescence Spectroscopy

Z. Wang

Photoluminescence is a process in which a substance absorbs photons and then re-emits

photons (usually at somewhat lower energies with smaller number of photons).

Photoluminescence spectroscopy analyzes the distribution of energies involved in the

photoabsorption and photoemission processes, the efﬁciency of the photoemission as well as

their temporal characteristics. Because of the non-destructive and contactless nature,

photoluminescence spectroscopy applies to solids, solution, solid suspensions, and gaseous

materials, making it a highly versatile and sensitive technique for molecular detection and

structural analysis.

16.3.5.1 The Physical Principle of Photoluminescence Spectroscopy

The physical principle of photoluminescence spectroscopy is depicted by the Jablonski

diagram [57] (Figure 16.19), named after the Polish physicist Aleksander Jabło

nski.

Absorption of photons promotes an electron from the electronic ground state (S

) to certain

vibrational levels (v =0,1,2,...), of the ﬁrst (S

) or higher electronic excited state (S

2,3

...).

The excited electron quickly relaxes to the lowest vibrational level (v = 0) of that excited state

(typically within 1 ×10

−12

s) by vibrational relaxation (VR), then relaxes to S

through

internal conversion (IC), which occurs when a vibrational state of an electronically excited

state can couple to a vibrational state of a lower electronic state. The resulting electron can

further deactivate either radiatively by emitting a photon (FL) or non-radiatively through one

of three processes: (1) IC quenching to the ground state; (2) collisional quenching (CQ) to the

ground state; or (3) intersystem crossing (ISC) to a triplet excited states (T

) that usually lies at

lower energy relative to S

. From were, the electron can either emit a photon by

phosphorescence or deactivate through IC to S